Why Labour’s poll lead really isn’t very significant

Anyone with even the smallest interest in political science should know that it’s important not to get too excited about a single opinion poll, no matter how extraordinary the finding. In fact, even more caution should be applied if the finding is rather dramatic, as it seems to be with the latest YouGov poll. It shows Labour ahead for the first time since January, by a mere 2 per cent at 35 per cent, and now ahead of the Tories, who have dropped 5 points since a fortnight ago, to 33 per cent. So it is the first and only Labour lead since before the vaccine rollout and when disillusion with the government’s handling on the pandemic was deepening into a crisis of confidence. Now, though, just when the vaccine is supposed to have created a wall of protection against Covid and things are supposed to be normalising, we discover that the government has suffered a sharp drop in support. It seems to be down to the recent tax hikes to pay for health and social care, perhaps with some growing misgivings about the persistence of Covid, shortages and the withdrawal from Afghanistan.

It’s tempting, just to see what this might translate into at a general election. Labour supporters should look away now. Their poll rating, if repeated at a general election would still be one of the lowest in such a contest in the party’s history, and would leave Starmer with 283 seats in parliament, a gain of 80, but still some 43 short of an overall majority – a measure of how big a mountain Labour faces. Boris Johnson would have 268 MPs, down 97, and the Liberal Democrat’s would be boosted by 8 to 19, still some way short of the peaks of popularity they enjoyed under Charles Kennedy and Nick Clegg. Labour would be able to govern with the help of the SNP and/or the Liberal Democrats and the sole Green MP. The Conservatives would need the help of Ulster Unionists, a bit of a “big ask” in light of the Northern Ireland protocol Mr Johnson signed up to.

The optimistic case for Johnson and his party is that the electorate is volatile and might forget about the current tax hikes by the next election. We are also still in a pandemic, and to be only a couple of points short of the opposition midway through a parliament even in normal times would usually be counted a triumph. There is still all to play for in an election that might be delayed, in extremis, until the end of 2024. What’s more, the repeal of the Fixed Term Parliament Act will allow Johnson to choose the optimal moment for a poll, there may be some favourable boundary revisions, the new voter identification rules will probably favour Tory candidates, and the economy and wage levels may improve faster and more sustainably than people assume. If so, then the Tories should be head again in due course. At the moment, some Tory support seems to be seeping away to the Reform Party, rather than straight to Labour.

Indeed, that is the worry for Labour strategists. Arguably, Labour support ought to be way into the 40s now in percentage terms, and the party on track for a healthy majority. That is the criticism levelled at the current leadership both by (for want of better labels) Blairites and Corbynites. Instead it seems to the Lib Dems and the Reform Party (formerly the Brexit Party) who are the repositories of protest. It is fair to add, though, that the present poll reading would represent about a 6 per cent swing from Conservative to Labour, high by most historical standards, but some way short of Blair (8.8 per cent) in 1997 and Corbyn (9.6 per cent) in 2017.

So, for both parties the glass is half full and half empty, but perhaps trending more on the half-empty mark. It is hard for anyone to claim much of a democratic mandate on a third of the votes cast (around a quarter of the electorate or less), and unable to command an overall majority in the Commons. As things stand, taking a broader view of the run of polls this year, it’s fair to conclude that both main parties of government (excluding the SNP in Scotland) are failing to impress the electorate.